Amazon QuickSight Now Generally Available
aws.amazon.com
aws.amazon.com
As a software engineer, I have been in a tricky situation a couple times when my boss asked me to build out data visualization features in our internal admin dashboards--including generic visualization of arbitrary data in our SQL database.
Of course, the correct answer is, "We have better things to use our time on, we should buy something for this". But the boss is never happy to hear that:
* These services start at around $1000/month, even for very small-scale data.
* There are additional charges per-user.
* There are additional charges (hundreds per month) for more data sources (e.g. if you have a second SQL database, or dump some data to Redshift)
What's more, it is next to impossible to find pricing information on these service's web sites. Instead you resort to Quora[0] and hope the information is accurate.
I have used both Periscope and Looker, and loved them. But according to the Quicksight pricing page[1], my company could get everything it needs from Quicksight for under $50/month.
After I hit submit I am planning to run to my boss and tell him I finally found the one. (Of course, we'll see how the demo goes first ;)
[0] https://www.quora.com/Can-anyone-share-their-experience-and-...
I'm very hopeful for quicksight, but unfortunately it feels very clunky and hard to use at the moment. Hopefully they improve the clunky UI and give it a SQL inerface.
QuickSight is going to be leveraged for more than just data analysis. They'll (probably) be using this for cost exploring and other features down the road (ELB Log Analysis, VPC Flow Log Analysis).
There are several limitations that I ran into. If you want to use a Redshift cluster as a data source it has to be a) publicly accesable and b) you have to whitelist the QuickSight IP range. QuickSight is launched in Amazon's infrastructure (their account and VPC) and that places limitations on what can be accessed (no VPC peering).
As an operations engineer, my biggest feedback was "let me launch my own QuickSight resources in my own VPC". Don't get me wrong, the fact that you get an AD instance, SPICE DB, and Web UI at a button click is nice, but I want to have more freedom to control and secure it. Just my two cents.
How did you find QuickSight's IP range? I saw it referenced in the FAQ, but couldn't find the actual IP itself anywhere.
http://docs.aws.amazon.com/quicksight/latest/user/configure-...
I opened a support ticket last night after being unable to find this exact piece of information. If only there were a way to make a web page reference some other web page...
Oh gods, this day has come -- Tableau is now a traditional BI company. I remember the days when traditional BI meant BusinessObjects, Cognos or Crystal Reports.
Step 1: develop a Trojan horse to gather growth and profitability data (QuickSight).
Step 2: roll out home-grown apps to compete with their most profitable AWS customers.
This is exactly how Amazon.com operates[1].
I can only imagine how the conversation went:
Bob: We need to diversify!
Earl: You're right! Let's do some research.
Susan: That'll cost a ton...
Ed: Eureka! We'll just let everyone give us their data.
It'll be a quick sight into all the markets!
Becky: Ed... You're a genius!
Ed: I am?
Becky: Yes. We'll call it QuickSight!
Needless to say that doesn't work, especially given the number of enforcement officials at the SEC and the capacity there, and the fact that Apple would have to sue it's own employees (plus the fact that you wouldn't want them to be heavy-handed at all).
So SOX means that very few employees can have access to such data. The more enterprising amongst them still do of course - as do the ones that you really don't want to have access to it. The ones that would both "need" the data (not really), and would know exactly how to profit from it - general and financial management.
https://aws.amazon.com/compliance/data-privacy-faq/
The reason AWS is successful is because Amazon takes privacy and security seriously. Why does Netflix trust AWS when there's Prime Video? Because Netflix knows that they have legal recourse against Amazon if they tried to do anything underhanded. I suspect Amazon would quickly fire anyone who breached customer data.
The logic must be: more money in apps than infrastructure.
Opposite of Google, they have the apps and now want the infrastructure too - I think it's a mistake for Amazon to go down this route, there is too much competition in this space and so far playing with QuickSight I don't see anything new/different from other analytics apps available on Amazon marketplace [1].
It feels like an experiment vs. a new direction for the company
[1] https://aws.amazon.com/marketplace/b/2649336011?ref_=gtw_nav...
if you want api-level analytics stuff and dashboards, have a look at amazon mobile analytics. it's an interesting alternative to google analytics that lets you control the session key (so it's easy to correlate events from front-end and back-end), and the event data is also yours (so it can easily be exported).
PaaS is an important market, and it'd be suicidal for AWS to not leverage its IaaS offerings and build platforms on top. This is kind of what Google does with its apps on Android. They have a replacement for every popular app out there that isn't a game.
In this app economy, something like Google Firebase is going to be a huge player if isn't already and that's the segment you can't compete with if you don't have PaaS offerings.
With Google controlling the entirety of developer ecosystem on Android which is the most dominant phone OS by far via Google Play, it is paramount that AWS doesn't turn a blind eye to it.
In a way, GCP might well be the most widely used Cloud Platform out there by virtue of Google Play integrations in billions of Android Phones, we are choosing to ignore that fact at our own peril.
AWS packages itself as a great enterprise offering. But GCP cloud take over business that are mobile first (Snapchat, for instance), I'd say, if competitors like AWS stop innovating and building.
AWS generally tends to complement its offerings such that they can avoid most of existing users of one or more services from looking at other alternates.
I'm not familiar enough with all those alternatives you cited to make a feature-by-feature comparison but one key difference is those other tools all use AMI (vm images). The Amazon QuickSight doesn't seem to need a deployment and management of any AMI. The customer just uploads a csv or Excel or points at a data source like Salesforce and instantly starts slicing & dicing the data. QuickSight seems to be positioned similar to Google's Data Studio offering.
>The logic must be: more money in apps than infrastructure.
That seems inevitable. Amazon AWS proved that IaaS being higher up the value chain than bare-metal services was more profitable. Rackspace's decline was proof of that. To continue the climb up the value chain, everybody including Google Cloud and MS Azure is offering higher valued-added services. Even a company like Github couldn't survive purely on on hosted Git disk space. (A "dumb" data pipe and disk space to a shared git repo.) They are moving up the value chain to build out full life-cycle project management to compete with MS Visual Team in the cloud.
This, on first look, definitely looks like it will be simple enough to have our operations team work directly. That's quite exciting!
My info's in my profile, I always love to meet tech folk working in Nairobi.
I evaluated trial versions of the following tools: Looker, Periscope, Mode, Tableau, QuickSight, and Microsoft PowerBI. They are all very similar in my opinion and at the end of the day will all provide the ability to view dashboards and reports through a web interface. The tangible differences come down to how exactly data refresh happens and the customizability of dashboards and reports. Each tool also offers varying degrees of useful edge features like sharing, alerting, privileges, and saved queries. I eventually settled on Microsoft PowerBI.
The biggest advantage for using PowerBI in my mind is the degree to which reports can be customized. Compared to QuickSight, PowerBI is leagues ahead in terms of customization, while still providing enough predefined templates that you are not constantly thinking about trivial styling choices. More than any other feature, Microsoft gets this right and executes on it better than any other BI tool I've tested. There is even a growing gallery of community visualizations that can be used for free (https://app.powerbi.com/visuals/), and you can write your own visualization templates using TypeScript.
I've been using PowerBI for several months now, and each month Microsoft releases an update that targets the biggest frustrations I've had, such as lack of pre-defined data connections or pre-defined styles. It's still missing some key features I would like (different forms of aggregations, etc) but it's still better than QuickSight, no question. As far as pricing goes PowerBI actually starts with a gratuitous free tier. The free part ends when you want to enable the "groups" feature, which allows you to set up privileges for which users can view certain dashboards. The paid version also allows more data to be stored in the SaaS. You need a premium account which is $10/per user/per month. The pricing + the speed of creating and uploading reports made PowerBI the winning tool. So far the team has been very happy with it.
So it begins. How sticky are customers in this space to existing solutions?
It says I can upload a CSV. What would be in the CSV? One row per sale, for instance? Does it then do a bunch of magic to, among other things, trend that pile of data?
Thanks in advance!
Disclaimer: I work at https://vida.io, an open source visualization tool.
The idea: Re-sell pre-built dashboards and reports for specific niches.
Of course, you can roll your own using various javascript graphing libraries but it would be nice if everything was taken care of and one could just build the various reports and visualizations.
Of note (in case anyone is interested) was this open source project which appears to be a complete solution: https://github.com/piwik/piwik
The page feels very sluggish on a i7, 16GB MBP. What machines are they using to test it?
The Google Data Studio is only for dashboards (i.e. static visualizations) these other tools offers standard BI operations like rollup, drillup, pivoting and etc.
As others have stated, this offering directly competes with Google's Data Studio which was recently just announced.